3.1. Principle of the t-PER-based pathogen nucleic acids biosensor
The principle of the proposed biosensor based on t-PER is shown in Scheme 1. The human papillomavirus (HPV) was chosen as the model. Two reactions are consisted in this system, including toehold mediated-PER amplification and signal out. The whole PER system includes hairpin with stop site, primer, and Bst DNA polymerase. The 3’ end of hairpin and protector strand were modified by inverted T base,which can inhibit the self-extended by DNA polymerase. The stop site is consisted of consecutive three G bases, which can prevent the polymerization in the presence of substrate (dATP, dCTP, and dTTP). At first, the hairpin was blocked by the protector strand, which leads to inactivation of DNA polymerase. In the addition of target, the blocked hairpin was released by toehold-mediated strand displacement. The primer was combined with the hairpin, and extended by the active DNA polymerase to the stop site, which is. Because of the breathe reaction and repeat sequence “a” design of hairpin stem, the extended primer would be set back to the initial binding site of hairpin, and the DNA polymerase works again. Then, a primer exchange reaction is completed. After multiple exchanges reacted, long massive DNA products were obtained. It would open plentiful molecular beacon (MB) to realize recovery of FAM fluorescence. The whole system was incubated at 37℃ in one pot. Thus, a simple and specific fluorescence platform was developed for HPV detection.
3.2. Characterization and feasibility of the proposed biosensor.
The polyacrylamide gel electrophoresis was first used to demonstrate the feasibility of the t-PER. As shown in Fig. 1a, single sequences (Lane 1: hairpin; Lane 2: protector) and double-strand products (Lane 3: target + protector; Lane 4: protector + hairpin) were synthesized without other heterochains. The protector could be hybridized with the target successfully by the toehold-mediated strand displacement (Lane 5). The PER was initiated to generate long single-strand products (Lane 6). In the presence of target DNA, the t-PER system could be realized as designed (Lane 8). Without the addition of target, no high molecular weight bands were obtained (Lane 7), showing an outstanding stability of this system.
The feasibility of this fluorescence platform based t-PER was further studied by measuring the fluorescence signals of recovered FAM. As described in Fig. 1b, when the hairpin and primer mixed with denatured DNA polymerase, the fluorescence intensity could hardly be observed (curve d). On the contrary, the PER was successful initiated by adding the activated DNA polymerase. Then, the multiple long DNA products obtained from PER would open the MB, which generate much higher fluorescence intensity (curve a). With the addition of protector, the PER will be blocked lead to weak fluorescence signal (curve c). In the presence of target DNA, the blocked PER can be triggered again (curve b), indicating that the target initiated PER amplification by toehold-mediated strand displacement. These results demonstrated that the t-PER could specificity recognize the target and generate an amplified fluorescence signal.
Most isothermal amplification for pathogen nucleic acids detection includes two steps (target specific recognition and signal amplification), which lead to the cumbersome operation[38]. To study the feasibility of the one-step reaction based t-PER, the fluorescence measurements of the one-step and two-step processes were adopted, respectively (Fig. 1c). The fluorescence signal of one-step process was 5.3% higher than that of two-step process, which proven that it is feasible to perform the toehold-mediated displacement and trigger PER amplification simultaneously, and also efficient for analytical performance. This may be due to the fact that when the target initiated the strand replacement and the PER system was separated for continuous reaction, there is relatively sufficient time for primer to generate the long single DNA strand, which tosome extent hampered the emergence of hairpin-protector-target terpolymer. While the concurrent occurrence of the two reactions made the target could bind with the protector-hairpin immediately. In addition, the fluorescence kinetic monitoring was used to further verify the viability of the one-step reaction. The results exhibited a time-dependent fluorescence intensities increase until a plateau from different target concentrations within 30 min (Fig. 1d). These above results demonstrate that the designed t-PER method could be adopted for the quantitative detection of pathogen nucleic acid in one-step.
3.2. Optimization of reaction conditions
To obtain optimal analytical performance of the proposed platform, several important parameters were optimized. Firstly, the concentration of DNA polymerase was investigated. As shown in Fig. 3a, the signal to noise ratio (S/B) achieved the perk value at a concentration of 0.2 U/µL DNA polymerase, thus the 0.2 U/µL was selected as optimal concentration of DNA polymerase for subsequent experiments. Next, the length of primer probe from 8 to 16 nt was optimized. As depicted in Fig. 3b, the S/B increased as the increasing length of the primer up to 12 nt, the S/B decreased with the further increasing the primer length, demonstrating that increasing the length of primer not only improved the efficiency of t-PER amplification but also enhanced background signal. The effects of reaction time and temperature on the proposed platform were also studied. Figure 3c showed that the S/N reached the maximum at 60 min of incubating time, followed decline with the increasing time. As shown in Fig. 3d, the S/N reached the maximum at a temperature of 37°C. Thus, the incubating time of 60 min and reaction temperature of 37°C were chosen for the following test.
3.3. Analytical performance of the proposed platform.
The sensitivity of this method was studied first using synthesized target DNA with varying concentrations. As described in Fig. 4a, the results shown that the fluorescence intensity (FI) increased with the increasing of target DNA concentration (Ctarget) from 100 fM to 50 nM. There was a good linear correlation between the current signal and the logarithm of Ctarget in the range from 100 fM to 1 nM (Fig. 4b), fitted as FI = 7.346 lgCtarget (fM) + 28.826 with correlation coefficient of 0.999. The limit of detection (LOD) was 18 fM based on 3SD corresponding to the blank tests. Comparing with other fluorescent or colorimetric sensors based on isothermal amplification (Table S2), this method has moderate sensitivity. Notably, it allows detection to complete within 60 min in one step, without additional labeling or washing processes.
The specificity of this platform was evaluated by detecting multiple analogous sequences, including double-base mismatched DNA (DM), four-bases mismatched DNA (FM), random DNA (RS), five HPV subtypes (HPV 06, 18, 31, 33, and 58) and two human-associated viruses (HBV and EBV). As shown in Fig. 4c, the FI from the target DNA was 4 times high than that of DM and over 10 times higher than that of other analogous sequences, exhibiting high specificity of the developed platform. To further demonstrate the specificity of t-PER, we designed corresponding toehold-mediated strand replacement for specific distinguishing HPV subtypes (HPV 06, 18, 31, 33, and 58). As shown in Fig. 4d, compared to target HPV subtypes, the platform produced weak FI when responded to other HPV subtypes. The reproducibility of this biosensor was studied by measuring the samples contained 10 pM target DNA in six different batches. The variable coefficient of 2.8% was obtained, exhibiting a decent reproducibility of this platform.
3.3 Target selection in HPV whole genome.
Although some toe-hold based methods can be used in broad-spectrum pathogen detection, their clinical application is still limited. Some studies have pointed out that toehold-mediated recognition is restricted by the folding format of the gene region. The selection of target is important for the clinical application of toehold based methods[39, 40]. Therefore, we tested three specific target site using t-PER for HPV detection, including capsid protein coding gene (CPCG), L1 gene, and L2 gene (Fig. 2a). Before the performance analysis on this method for the whole genome, we adopted HPV DNA extracted from clinical positive sample, which confirmed by RT-PCR (Ct value: 23.1). As illustrated in Fig. 2b-d, upon the target addition the t-PER reaction showed fluorescence signal enhancement, meaning that all of these three target genes could successfully work to transform the whole genome to the corresponding hairpin-primer template and further PER amplification. Notably, compared to the L1 gene (Fig. 2b) and L2 gene (Fig. 2c), the t-PER assay on the CPCG (Fig. 2d) showed better analytical performance (stronger fluorescence signal and higher signal to noise ratio (Fig. 2e)). The potential reason might be that the unexpected three-dimensional conformation and self-folding of the whole genome inhibits the correct binding of protector to corresponding hybrid site. To further explain the high efficiency of t-PER for CPCG, we carried out DNA structure analysis using Nupack[41]. By comparing the folded-binding sites with an additional 30 nt and 60 nt up and downstream, the L1 gene and L2 gene folded more compact than CPCG. And the whole hybridize site of L1 and L2 was in a structured region, whereas for the CPCG, the toehold region of the hybridize site was in a high probable unpaired region (Fig. S1). These results are consistent with the results of fluorescence measurement. The above results reveal that it is vital to choose a target sequence within the whole genome.
3.3 Clinical utility of the proposed platform
The clinical utility of this platform was studied by detecting HPV infections in clinical cervical swabs samples. As t-PER works through direct target binding, rather than via target amplification, we speculate that the t-PER could have reformative compatibility for clinical samples Fig. 5a. To achieve direct detection in clinical cervical swabs samples, we proposed a short heat lysis (95°C, 5 min) for released viral DNA (Fig. 5b) in five clinical positive samples. As shown in Figure S2, the FI obtained from heat lysis based t-PER was consistent with that from DNA extraction, the variable coefficient was under 2.5%. These results demonstrated that the viral DNA was released successfully, and its integrity was preserved. These may attribute to following reasons : 1) t-PER process through direct target replacement to initiate a polymerase. The polymerase-based elongation is hardly disrupted. 2) similar to biological functions, the design of t-PER has improved compatibility between enzymes and DNA strands. Then, we adopted the t-PER assay to determination DNA targets directly in these cervical swab samples (including 16 positive samples and 8 negative samples). It is worth mentioned that this platform distinguished all samples correctly. The results are consistent with those obtained from qPCR (Fig. 4g). Moreover, the signals of positive samples were significantly higher than those of negative samples (*P < 0.005, Student’s t test), while Ct values of positive samples range from 22.3 to 32.03(Figure S3, Table S3).
3.5. The proposed platform used for other pathogen virus detection.
To better demonstrate the clinical utility, the developed platform was also used to detect HBV and EBV, which are vital biomarkers for hepatitis monitoring and nasopharyngeal carcinoma diagnosis, respectively. At first, the corresponding toehold-design and target region screening in genome was implemented (Fig. 5a and c). Considering that EBV LMP-2A gene and HBV S gene are widely adapted for clinical application, the corresponding toehold-mediated strand placements were designed to identify these regions (Figure S4). As shown in Fig. 5C and D, this platform could detect all samples (including 10 positive samples and 5 negative samples) infected HBV and EBV with 100% concordance of the results obtained from qPCR. Significantly, the relative fluorescence signals of positive samples were highly consistent with the Ct values (Table S4). In order to further study the versatility of the proposed method, the RNA target of the Ureaplasma Urealyticum (UU) in 10 clinical samples (including 6 positive samples and 4 negative samples) were also successfully tested. These results also demonstrated that this method could be used for RNA target detection (Figure S5 and Table S5).